From: dcm3c@xxxxxxx
Reply-To: femm@xxxxxxxxxxx
To: femm@xxxxxxxxxxx
Subject: [femm] Re: various
Date: Mon, 22 May 2000 10:01:07 EDT
In a message dated 5/18/00 5:34:54 AM Eastern Daylight Time,
rlbrambilla@xxxxxxx writes:
> These are my suggestions that, I hope, require simple modifications.
>
> GLOBAL : a command for a custom dafault settings (units, materials
etc.).
> Actually I use a template file.
Including some sort of standard (but user-modifiable) template would be a
good idea.
> GEOMETRY : commands for inserting points and lines by keyboard. In
large
> packages there are tens of this commands, most of them never used. Only
two
> commands are strictly enough (but unavoidable) for the model precision
:
> Point(x,y), and Line(x1,y1,x2,y2).
You can do this already. In pretty much any situation where a specific
point
needs to be specified, you can hit the <tab> key. A dialog will pop up,
allowing you to enter the coordinates of a point by keyboard, rather than
having to use the mouse.
> RESULTS : similar Point and Line keyboard commands to read results and
to
> paste values to the clip-board for exporting. Actually it is possible
to
> paste only the field picture, but I need to export values to other
programs
> for successive elaborations.
> Fine will be to have command to select the quantity I want the contour
> lines (as in actual xy-plot).
You can get the data from an xy plot to disk by checking the "write data to
text file" check box in the dialog where you select what sort of xy plot is
to be created. Having an option to send it to the clipboard might be a
good
idea as well.
> And now a little remark.
> In frequency dependent problems, inserting point currents,
> I insert peak value, i.e. Io for a varying current Io exp(jwt).
> Reading the results for a selected point I find Bx, By, |B|.
> I suppose again Bx exp(jwt), By exp(jwt).
> Now your |B|=sqrt(|Bx|^2+|By|^2) and the time factor is cancelled
> by conjugation. So it may be considered an "istantaneous constant
modulus".
> Usually we need the time averaged over a period so this value
> has to be divided by sqrt(2). To avoid this external operation
> I insert Ieff=Io/sqrt(2) as a point current. But in this case
> I cannot suppose Ieff exp(jwt)! So it seems to me that the output
> <|B|> = |B|/sqrt(2) and peak value insertion would be more clear.
In an attempt to be consistent and make things flow sort of "naturally"
from
the magnetostatic to harmonic cases, I have kept all results in terms of
amplitudes, rather than RMS quantities. I might see about putting in a
choice of "modes" for the output results, either the complex amplitudes (as
it does now), amplitude and phase angle, or RMS quantities.
> Hoping for future improvements,
> sincerly yours Roberto
Thanks for your comments.
-----------------------------------------
In a message dated 5/18/00 8:16:04 AM Eastern Daylight Time,
eckart.nipp@xxxxxxxx writes:
> A difficulty is the rectangular selection tool. With that it is
comparably
> troublesome to select e.g. a rotor of an asynchronous motor with all
its
> slots but not any part of the stator. This is particularly hard if the
> airgap is short. Maybe an optional circular selection tool can be
> developed. Also a deselection possibility, e.g. by holding shift, would
be
> useful. When playing around with a not predefined geometry also the
> possibility of moving parts per mouse may be useful.
Different selection tools might be useful. With the way that the selection
tool currently works, you can use it multiple times to select all the parts
of a geometry that has a more complicated shape--the previously selected
stuff doesn't automatically get deselected.
Anyhow, a good way to analyze a rotor at different positions in the 3.0
version is to assign all of the parts of the rotor to one "Group" as you
build it. Then, you can select the entire rotor with just one right-click
of
the mouse when the editor is in "group" mode.
> Can the circuit property also be used to define the voltage of an
exciter
> coil which is aymmetrically placed so that it is unknown in the
beginning
> how the total voltage distributes on forward and backward pathes?
There's a feature that's sort of "hidden" having to do with circuit
results.
In the postprocessor, choose the View | Circuit Props selection off of the
main menu. This allows you to see what voltage was required to impose a
prescribed current, or how much current was obtained from an applied
voltage
for a particular circuit (as appropriate). This might allow you to get the
information that you are after.
> Now I think I also found a bug in the group copy tool. When I chose a
> circle consiting of two 180deg arcs and copy it as a row, i.e. several
> copies at once, I get the desired circles but in addition a row of
anchor
> points.
I've been able to duplicate this bug--many thanks for catching it. This
was
the sort of thing that I was hoping to have people find with this beta
version....
> Greetings
> Eckart
Dave.
--
http://members.aol.com/dcm3c